25 years ago, Dr. Richard Lenski of Michigan State University began an ingenious experiment on E. coli bacteria that is still going and giving us a much deeper understanding of how evolution operates. Ars Technica has an article on the latest published findings, but let me first explain how the experiment was set up.
Lenski set up 12 different isolated environments (flasks, essentially) in his lab and put an equal amount of E. coli in each one, with the contents of those flasks being the same initially. Over time, he changed the environment in the flasks and let the bacteria reproduce. Every 500 generations, he removed a sample from each of the flasks and froze it, then let the rest continue to reproduce. Whenever something interesting happened — say, a big bloom in one of the flasks that indicates that something happened that facilitated greater reproduction — they could then take the current population, sequence the DNA and compare it to the frozen samples from that flask. This allows them to identify specific mutations that took place and when they took place (since they had samples from each flask every 500 generations).
One of the fascinating things that happened a few years ago is that one of the flask populations developed the ability to metabolize citrate (citric acid), meaning use it as food to fuel growth and reproduction. In the wild, E. coli can’t metabolize citrate if oxygen is present, but sub-population #3 in the lab had developed that ability, an obviously beneficial adaptation that would allow E. coli to survive and thrive in new environments when other sources of nutrients are not available. They were able to trace the development of that trait to a pair of mutations that took place at about 20,000 generations. Now the new findings:
Over the years, evolution has radically altered the bacteria. Some but not all the cultures evolved the ability to use citrate instead of glucose, leading to a significant growth advantage. Others picked up genetic changes that left them more susceptible to mutations, accelerating their ability to adapt. But these mutations took place tens of thousands of generations ago. Lenski decided to see if evolution was still driving increases in fitness, or if the bacteria had already achieved all the improvements they were likely to see.
To see if the tests were worthwhile, the researchers took the current group of bacteria (about 50,000 generations into the experiment) and compared them to some samples saved from generation 40,000. When competing for growth against each other in the same medium, generation 50K grew faster in nine of the 12 cultures, and it was roughly equal in the rest. To get a full fitness curve, Lenski’s group then set all 12 cultures into competitions with their own pasts, using as many as 40 competitions for each culture.
The end results for all cases were pretty clear: the bacteria made large gains in fitness during the first few thousand generations, after which the pace slowed. But the increase in fitness never really stopped. The fitness curves were tested to see if it looked like a hyperbola, which approaches but never exceeds a fitness limit, or a power law, in which case fitness will continue to increase over time even as its pace slows. The results were pretty clear: “the power law outperforms the hyperbolic model with a posterior odds ratio of ~30 million.”
And then this fascinating finding about the ability of natural selection to weed out harmful mutations and preserve beneficial ones:
The team then attempted to build a mathematical model that would describe the fitness increases based on the mutation rate. One of the notable features of their model is that it left out the occurrence of harmful mutations entirely, and it still produced accurate fits to the fitness curves. In other words, evolution seems to be so efficient at getting rid of harmful mutations that they can safely be ignored.
Another way this became apparent is the hypermutator strains. Half of the original dozen strains picked up mutations in genes that resulted in further mutations accumulating at a 100-fold higher pace. If negative mutations were much of a problem, then the hypermutators should show fitness curves that were below those of the remaining strains. Instead, they achieved higher levels of fitness than their peers by 10,000 generations and continued to stay ahead for over 40,000 generations.
This is science at its very best, designing a brilliantly conceived experiment that vastly increases our understanding of how nature operates.